Clinica Chimica Acta 430 (2014) 92–95
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
Peculiar observations in measuring testosterone in women treated with oral contraceptives supplemented with dehydroepiandrosterone (DHEA) Annemieke C. Heijboer a,⁎, Yvette Zimmerman b, Theo de Boer c,1, Herjan Coelingh Bennink b, Marinus A. Blankenstein a a b c
Department of Clinical Chemistry, VU University Medical Center Amsterdam, The Netherlands Pantarhei Bioscience BV, Zeist, The Netherlands QPS Netherlands BV, Groningen, The Netherlands
a r t i c l e
i n f o
Article history: Received 3 November 2013 Received in revised form 24 December 2013 Accepted 27 December 2013 Available online 7 January 2014 Keywords: Testosterone Assay Contraception SHBG Mass spectrometry DHEA
a b s t r a c t Total testosterone is considered to be decreased during the use of combined oral contraceptives. There is, however, considerable concern about the quality of testosterone assays, especially at low levels. We aimed to confirm testosterone levels measured by direct radioimmunoassay in a recent clinical trial with a state-of-theart LC–MSMS method. Surplus specimens with known testosterone levels collected during the study (Clinical Trial Registration number ISRCTN06414473) were reanalyzed with an LC–MSMS method. This method was compared to another LC–MSMS method that had shown to concur excellently to a reference method. Follow-up experiments were designed to explain the results. In contrast to our expectation, LC–MSMS measurements did not corroborate the data obtained by radioimmunoassay. Subsequent experiments showed that this could be attributed to a strong dependency of the radioimmunoassay on SHBG. Testosterone results (n = 198) obtained by direct radioimmunoassay showed a negative correlation to SHBG levels (r = −0.676; p b 0.001). By contrast, testosterone results obtained by LC–MSMS were not related to SHBG (r = 0.100; NS). In conclusion, our results indicate that total testosterone measurements during oral contraceptive use are unreliable when performed with assays sensitive to the SHBG concentration. The discrepancy with the literature can most likely be explained by the sensitivity of the immunoassay used to SHBG. Given the sharp increase in SHBG during the use of many oral contraceptives, total testosterone may not decrease, whereas its bioavailability, estimated by free testosterone levels, will be diminished. Studies aiming at restoration of testosterone homeostasis during oral contraception need to take this into account. © 2014 Elsevier B.V. All rights reserved.
1. Introduction It has been recognized for many years that use of combined oral contraceptives (COCs) lowers the concentrations of total and bioavailable testosterone [1]. A variety of undesired physiological effects has been associated with testosterone deficiency in postmenopausal women and it has been hypothesized that this might also apply to premenopausal women experiencing COC induced reduction in circulating testosterone. A recent study was aimed at restoring the decreased availably of testosterone (Clinical Trial Registration number Abbreviations: COCs, combined oral contraceptives; DHEA, dehydroepiandrosterone; DRSP, drospirenone; EE, ethinylestradiol; LC–MSMS, liquid chromatography–tandem mass spectrometry; RIA, radioimmunoassay. ⁎ Corresponding author at: Department of Clinical Chemistry, VU University Medical Center, P.O. Box 7057, 1007 MB Amsterdam, The Netherlands. Tel.: +31 20 4443872; fax: +31 20 4443895. E-mail address: [email protected] (A.C. Heijboer). 1 Present address: Analytical Biochemical Laboratory (ABL), Assen, The Netherlands. 0009-8981/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cca.2013.12.042
ISRCTN06414473; Zimmerman et al., submitted) by administration of dehydroepiandrosterone (DHEA) and, as expected, revealed lower total and free testosterone levels in the women taking COCs. It has also been known for a long time that measuring testosterone in sera from women presents a huge challenge to the laboratory, especially if the levels are low [2,3]. The need to introduce better testosterone assays has been emphasized by Rosner et al. [4]. In response to this broadly accepted appeal laboratories have introduced better assays and in this respect liquid chromatography tandem mass spectrometry (LC–MSMS) methods have been considered as the method of choice by a number of laboratories [5,6]. Similarly, manufacturers of immunoassays have made an effort to improve their assays [7]. A new reference management procedure for total testosterone has been proposed [8] and the Center for Disease Control has initiated a testosterone standardization project (http://www.cdc.gov/labstandards/pdf/hs/HoSt_ protocol.pdf, accessed August 2013). In the abovementioned trial on the restoration of testosterone levels during oral contraceptive use total testosterone was measured with a
A.C. Heijboer et al. / Clinica Chimica Acta 430 (2014) 92–95
direct competitive immunoassay and in view of the prevailing concern regarding the quality of testosterone assays in women, the present experiments were designed with the aim to confirm the data obtained with this immunoassay by LC–MSMS. 2. Materials and methods 2.1. Specimens Specimens were derived from a study designed to evaluate the effect of restoring the decreased availably of testosterone with daily administration of DHEA. Design and results of this study, Clinical Trial Registration number ISRCTN06414473, have been submitted for publication. Briefly, the study was a randomized, double blind, placebo-controlled, phase II study performed on 99 healthy new COC users at the University Hospital in Liège, Belgium. The intervention consisted of 3 cycles of a COC containing 30 μg ethinylestradiol (EE) and 3 mg drospirenone (DRSP) followed by 6 cycles in which the COC (Yasmin®) was combined with either 50 mg/d DHEA or placebo. Blood was taken in the morning after an overnight fast at the start of the study and after 3, 6 and 9 cycles of COC use. Serum was stored at −80 °C until analysis. 2.2. Testosterone and SHBG assays As part of the clinical study, testosterone levels were assessed by the Immuchem Double Antibody Testosterone 125I RIA kit according to the instructions of the manufacturer (ICN Biomedicals, MP Biomedicals, The Netherlands, RIAICN). Cross-reactivity with DHEA as reported by the manufacturer was b 0.01%. Two LC–MSMS methods have been applied to selected samples. LC– MSMSQPS was a fast analytical method involving a liquid–liquid extraction step using stable isotope-labeled testosterone (D3-testosterone) as the internal standard. The analytical procedure was as follows: 10 μL of an internal standard solution was added to 250 μL serum sample. After addition of 500 μL ammonium formate (0.1 M, pH 5.0) the mixture was extracted using 3 mL of a hexane–ethylacetate mixture (1:1). The tubes were vortexed and centrifuged for 4 min under ambient conditions at 1500 ×g after which the organic layer was separated and evaporated at 45 °C under a gentle stream of nitrogen. Five microliters was injected into an ultra-high-performance liquid chromatography system (Agilent 1290 series; Hewlett Packard, Palo Alto, CA) equipped with a Sciex API 4000 mass spectrometer (Applied Biosystems, Foster City, CA). The analytical column used was a hydrophilic interaction liquid chromatography column (Kinetex XB-C18 5 × 2.10 mm, inner diameter = 2.6 μm; Phenomenex, Torrance, CA). The following isocratic liquid chromatography solvent program was used: 50% mobile phase A [0.1% formic acid] and 50% mobile phase B [0.1% formic acid in acetonitrile]. The tandem mass spectrometry system was operated in positive ion mode. The retention time of testosterone and the internal standard was b1 min. The following multiple-reaction monitoring transitions were optimized: m/z 289.3 → 96.6 (testosterone) and m/z 292.3 → 96.6 (testosterone-d3). The calibration and quality control samples were prepared in steroid free serum (SeraCare, Milford, MA, USA). The calibration samples (n = 8) ranged from 0.0867 to 86.7 nmol/L (25.0 to 25000 pg/mL) and the quality control samples (at three concentration levels: 0.260 nmol/L (75.0 pg/mL), 4.33 nmol/L (1250 pg/mL) and 69.3 nmol/L (20 000 pg/mL)) were analyzed in duplicate. Intra-assay CV was 6.3%, 4.5%, and 3.3% at 0.26, 4.33 and 69.3 nmol/L (75.0, 1250 and 20000 pg/mL) respectively (n = 8). LC–MSMSVUmc was performed as described before, using D5testosterone as an internal standard [9–11]. Intra-assay CV was 9.5% and 6.0% at 0.21 and 1.98 nmol/L respectively (n = 53). In this method D3-testosterone and D5-testosterone have been shown to be equally efficient as internal standards [12]. Both LC–MSMS methods were shown to be free from interference from DHEA.
93
SHBG in the clinical trial was measured on the Roche Modular E170 platform, SHBGRoche, whereas in the present experiments the Architect immunoassay platform (Abbott diagnostics, Chicago, IL, USA) with an intra-assay CV of 6.3% and 4.9% at concentrations of 34 and 134 nmol/L respectively was used, SHBGAbbott. 3. Results The comparison between the results of testosterone measurements obtained during the trial with the RIAICN and the results obtained with LC–MSMSQPS is displayed in Fig. 1. When analyzed by LC–MSMS, total testosterone is not suppressed after three cycles of COC use. Both testosterone methods showed elevated testosterone levels following addition of DHEA to the COC (Fig. 1). Only in the samples obtained before the use of COC there was concurrence of the results of the testosterone assays. COC use resulted in lower testosterone levels when measured by RIA, irrespective of the addition of DHEA to the treatment. Subsequently, the LC–MSMSQPS was compared to LC–MSMSVUMC in 24 samples with testosterone values ranging from 0.53 to 5.12 nmol/L (by LC–MSMSQPS) selected from three subjects in each treatment arm taken after 0, 3, 6 and 9 treatment cycles respectively. The result (LC– MSMSQPS = 0.931 × LC–MSMSVUmc − 0.059; r = 0.9929; n = 24) establishes the excellent comparability of the two methods and, therefore, confirms the observation that, when measured with LC–MSMS the decrease in circulating testosterone that is observed with RIA is not confirmed. We hypothesized that insufficient release of testosterone from SHBG in the direct RIA procedure might be the cause of this apparent discrepancy. To test this hypothesis, testosterone was measured in 24 surplus serum specimens with a wide range of SHBG levels obtained from young women, 12 of whom were on different COCs. Specimens were analyzed for testosterone by RIAICN and LC–MSMSVUmc. The differences between the results are plotted in Fig. 2 and reveal a bias that is strongly dependent on the SHBGAbbott level (r = 0.79; p b 0.0001). Based on the results of this experiment, we compared the relationship between circulating total testosterone and SHBG using testosterone measurement by RIAICN and LC–MSMSQPS in the specimens from the trial. A statistically significant, negative correlation was found between testosterone measured by RIAICN and SHBGRoche (r = 0.676; p b 0.0001; n = 198). No such relationship was found between testosterone measured by LC–MSMSQPS and SHBGRoche (r = 0.100; NS; n = 196). The results in Fig. 3 confirm our suspicion that the results of the RIAICN are severely influenced by the concentration of SHBG in the specimen. As expected based on the good agreement between the LC–MSMS methods, testosterone measured by LC–MSMSVUmc was not related to SHBGAbbott (r = 0.062; NS; n = 24).
Fig. 1. Serum testosterone before and during treatment with combined oral contraceptives supplemented with DHEA or placebo after cycle# 3, as measured by radioimmunoassay (local RIA) and liquid chromatography–tandem mass spectrometry (LC–MSMS). Data are presented as mean ± SEM (n = 49).
94
A.C. Heijboer et al. / Clinica Chimica Acta 430 (2014) 92–95
Fig. 2. Differences between results of RIAICN and LC–MSMSVUMC for testosterone in relation to the SHBG concentration in the specimen. n = 24; r = 0.79; p b 0.0001.
4. Discussion The present study was designed to confirm the well documented observation [1] that administration of combined oral contraceptives causes a decrease in the circulating level of total testosterone. Worldwide concern about the quality of testosterone assays, especially at low levels, co-fostered our initiative. In sharp contrast to our expectation, however, measurements by LC–MSMS illustrated that after three months of use of 30 μg EE and 3 mg DRSP (Yasmin®), testosterone levels are indiscernible from those prior to the start of the medication. The LC–MSMS method used was compared to another, independent, LC–MSMS method, and these methods were found to agree well. This observation is in sharp contrast to the literature which describes, based on immunoassays with or without extraction [1], a significant decrease in both free and total testosterone during COC use. In contrast to testosterone, SHBG levels sharply rose during COC use in the trial from which the specimens were derived. After three COC cycles SHBG was 241% higher than before treatment and we hypothesized that this increase may have adversely influenced the results of the
Fig. 3. Total testosterone in relation to SHBG in samples from women before and after three cycles of combined oral contraceptive use, as measured by radioimmunoassay (RIAICN; panel A) or liquid chromatography–tandem mass spectrometry (LC–MSMSQPS, panel B).
radioimmunoassay used to assess testosterone levels. Indeed, the difference between the results of the RIA and LC–MSMS was shown to be strongly dependent on the SHBG concentration. This dependency may also explain the differences observed in testosterone levels following addition of DHEA to the COC (Fig. 1). We, therefore, suggest that, when measured with a state-of-the-art method, total testosterone does not change upon oral contraceptive use. Of course, there are differences in the design and quality of immunoassays [4] and not all immunoassays may be equally affected by the SHBG level in the specimens. One way to overcome the potential influence of SHBG on the test result is to extract testosterone with organic solvents prior to the immunoassay. In theory immunoassays preceded by extraction and chromatographic separation of the analyte should yield a result that shows more agreement with LC–MSMS than immunoassays without sample pretreatment. The effect of COC use on total testosterone assessed with extraction based methods should therefore be less than those assessed with direct immunoassays. In fact this was observed in the systematic review [1]. The supplemental material to that paper shows that COC use results in a 0.56 (95% CL 0.50–0.62) nmol/L decrease in serum testosterone when analyzed by direct methods (45 studies on 2098 patients), and 0.37 nmol/L (95% CL 0.26–0.47) when analyzed by methods using extraction and chromatography (17 studies on 522 patients). The difference was found to be statistically significant, p = 0.003 but the observation that even with extraction methods a decrease in total testosterone is observed on meta-analysis requires further consideration. The individual studies collected in the meta-analysis differ with respect to the conclusions. Thirty-five of the 45 studies using direct assays, 78%, report a significant testosterone decrease in the meta-analysis, versus only 9 of the 17 studies with extraction assays (55%). That not all extraction based assays fail to show a significant decrease may have to be attributed to methodological influences and may require further investigation. This may, however, be hindered by the fact that many of these methods may no longer be in use. Interference from the components of the COCs used is considered unlikely, as this would have led to apparently increased testosterone results in the competitive immunoassays and not to a decrease. Similarly, insufficient extraction recovery is also not likely to explain the phenomenon as this would equally affect specimens from control and COC cycles. The only exception to this would be if the increased SHBG in the specimens obtained during COC use would cause a difference in extraction efficiency. Dependency on the concentration of binding proteins is not unique to testosterone, since it has also been reported for vitamin D [13]. Unfortunately, like immunoassays, LC–MSMS methods also do not all yield the same result, especially at low levels [14,15] and with some LC–MSMS methods, total testosterone has been reported to decrease modestly during oral contraceptive use [16]. Moreover, these authors report on age-specific reference values tor testosterone in women, whereas we have not taken possible age differences into account. As in our study however, free testosterone was much more suppressed than total testosterone. Based on their results on postmenopausal women, Legro and colleagues [15] claim that the radioimmunoassays, when properly maintained, should yield comparable results to LC–MSMS methods. This may be true, but in postmenopause SHBG levels are not that high compared to after use of some oral contraceptives. Provided that their RIA can be shown to be free from the SHBG interference as described here, we agree with their conclusion. Our finding that, when measured by LC–MSMS, total testosterone, in contrast to earlier studies, does not change during COC use has no impact on the notion that women on COCs may experience androgen deficiency. The substantial increase in SHBG concentration during the use of many COCs will lead to a sharply decreased free testosterone level which may be responsible for the adverse effects reported. Studies on the benefit of maintaining bioavailable testosterone to the pre-COC level will, therefore, have to include assessment of SHBG and a proper method of measuring or calculating free testosterone. Preferred
A.C. Heijboer et al. / Clinica Chimica Acta 430 (2014) 92–95
methods are the mass action equation and equilibrium dialysis [3], the latter being quite laborious though. In summary then, in contrast to a vast number of reports in the literature, our results support the hypothesis that total testosterone concentrations do not change during the use of combined oral contraceptives when measured with a state-of-the-art method. The discrepancy with the literature can most likely be explained by the sensitivity of immunoassays to SHBG, as we showed strong dependency on SHBG concentrations in a testosterone immunoassay. Obviously, our data are limited to one immunoassay and the SHBG dependency of other immunoassays should be documented. Acknowledgments The authors gratefully acknowledge the expert technical assistance by F. Martens (VUmc) and B. Ottjes (QPS). References [1] Zimmerman Y, Eijkemans MJC, Coelingh Bennink HJT, Blankenstein MA, Fauser BCJM. The effect of combined oral contraception on testosterone levels in healthy women: a systematic review and meta-analysis. Hum Reprod Updat 2014;20:76–105. [2] Taieb J, Mathian B, Millot F, et al. Testosterone measured by 10 immunoassays and by isotope-dilution gas chromatography–mass spectrometry in sera from 116 men, women, and children. Clin Chem 2003;49:1381–95. [3] Miller KK, Rosner W, Lee H, et al. Measurement of free testosterone in normal women and women with androgen deficiency: comparison of methods. J Clin Endocrinol Metab 2004;89:525–33. [4] Rosner W, Auchus RJ, Azziz R, Sluss PM, Raff H. Position statement: utility, limitations, and pitfalls in measuring testosterone: an Endocrine Society position statement. J Clin Endocrinol Metab 2007;92:405–13.
95
[5] Soldin SJ, Soldin OP. Steroid hormone analysis by tandem mass spectrometry. Clin Chem 2007;55:1061–6. [6] Stanczyk FZ, Clarke NJ. Advantages and challenges of mass spectrometry assays for steroid hormones. J Steroid Biochem Mol Biol 2010;121:491–5. [7] Groenestege WM, Bui HN, ten Kate J, et al. Accuracy of first and second generation testosterone assays and improvement through sample extraction. Clin Chem 2012;58:1154–6. [8] Botelho JC, Shacklady C, Cooper HC, et al. Isotope-dilution liquid chromatography– tandem mass spectrometry candidate reference method for total testosterone in human serum. Clin Chem 2013;59:372–80. [9] Bui HN, Struys EA, Martens F, et al. Serum testosterone levels measured by isotope dilution-liquid chromatography–tandem mass spectrometry in postmenopausal women versus those in women who underwent bilateral oophorectomy. Ann Clin Biochem 2010;47:248–52. [10] Bui HN, Sluss PM, Blincko S, Knol DL, Blankenstein MA, Heijboer AC. Dynamics of serum testosterone during the menstrual cycle evaluated by daily measurements with an ID-LC–MS/MS method and a 2nd generation automated immunoassay. Steroids 2013;78:96–101. [11] Bui HN, Schagen SE, Klink DT, de Waal HA Delemarre-van, Blankenstein MA, Heijboer AC. Salivary testosterone in female-to-male transgender adolescents during treatment with intra-muscular injectable testosterone esters. Steroids 2013;78:91–5. [12] Bui HN, Blankenstein MA, Heijboer AC. Isotopically labelled testosterone derivatives as internal standards in liquid chromatography–tandem mass spectrometry. Ann Clin Biochem 2013;50:275–6. [13] Heijboer AC, Blankenstein MA, Kema IP, Buijs MM. Accuracy of 6 routine 25hydroxyvitamin D assays: influence of vitamin D binding protein concentration. Clin Chem 2012;58:543–8. [14] Thienpont LM, Van Uytfanghe K, Blincko S, et al. State-of-the-art of serum testosterone measurement by isotope dilution-liquid chromatography–tandem mass spectrometry. Clin Chem 2008;54:1290–7. [15] Legro RS, Schlaff WD, Diamond MP, et al. Total testosterone assays in women with polycystic ovary syndrome: precision and correlation with hirsutism. J Clin Endocrinol Metab 2010;95:5305–13. [16] Haring R, Hannemann A, John U, et al. Age-specific reference ranges for serum testosterone and androstenedione concentrations in women measured by liquid chromatography–tandem mass spectrometry. J Clin Endocrinol Metab 2012;97:408–15.
Contents lists available at ScienceDirect
Clinica Chimica Acta journal homepage: www.elsevier.com/locate/clinchim
Peculiar observations in measuring testosterone in women treated with oral contraceptives supplemented with dehydroepiandrosterone (DHEA) Annemieke C. Heijboer a,⁎, Yvette Zimmerman b, Theo de Boer c,1, Herjan Coelingh Bennink b, Marinus A. Blankenstein a a b c
Department of Clinical Chemistry, VU University Medical Center Amsterdam, The Netherlands Pantarhei Bioscience BV, Zeist, The Netherlands QPS Netherlands BV, Groningen, The Netherlands
a r t i c l e
i n f o
Article history: Received 3 November 2013 Received in revised form 24 December 2013 Accepted 27 December 2013 Available online 7 January 2014 Keywords: Testosterone Assay Contraception SHBG Mass spectrometry DHEA
a b s t r a c t Total testosterone is considered to be decreased during the use of combined oral contraceptives. There is, however, considerable concern about the quality of testosterone assays, especially at low levels. We aimed to confirm testosterone levels measured by direct radioimmunoassay in a recent clinical trial with a state-of-theart LC–MSMS method. Surplus specimens with known testosterone levels collected during the study (Clinical Trial Registration number ISRCTN06414473) were reanalyzed with an LC–MSMS method. This method was compared to another LC–MSMS method that had shown to concur excellently to a reference method. Follow-up experiments were designed to explain the results. In contrast to our expectation, LC–MSMS measurements did not corroborate the data obtained by radioimmunoassay. Subsequent experiments showed that this could be attributed to a strong dependency of the radioimmunoassay on SHBG. Testosterone results (n = 198) obtained by direct radioimmunoassay showed a negative correlation to SHBG levels (r = −0.676; p b 0.001). By contrast, testosterone results obtained by LC–MSMS were not related to SHBG (r = 0.100; NS). In conclusion, our results indicate that total testosterone measurements during oral contraceptive use are unreliable when performed with assays sensitive to the SHBG concentration. The discrepancy with the literature can most likely be explained by the sensitivity of the immunoassay used to SHBG. Given the sharp increase in SHBG during the use of many oral contraceptives, total testosterone may not decrease, whereas its bioavailability, estimated by free testosterone levels, will be diminished. Studies aiming at restoration of testosterone homeostasis during oral contraception need to take this into account. © 2014 Elsevier B.V. All rights reserved.
1. Introduction It has been recognized for many years that use of combined oral contraceptives (COCs) lowers the concentrations of total and bioavailable testosterone [1]. A variety of undesired physiological effects has been associated with testosterone deficiency in postmenopausal women and it has been hypothesized that this might also apply to premenopausal women experiencing COC induced reduction in circulating testosterone. A recent study was aimed at restoring the decreased availably of testosterone (Clinical Trial Registration number Abbreviations: COCs, combined oral contraceptives; DHEA, dehydroepiandrosterone; DRSP, drospirenone; EE, ethinylestradiol; LC–MSMS, liquid chromatography–tandem mass spectrometry; RIA, radioimmunoassay. ⁎ Corresponding author at: Department of Clinical Chemistry, VU University Medical Center, P.O. Box 7057, 1007 MB Amsterdam, The Netherlands. Tel.: +31 20 4443872; fax: +31 20 4443895. E-mail address: [email protected] (A.C. Heijboer). 1 Present address: Analytical Biochemical Laboratory (ABL), Assen, The Netherlands. 0009-8981/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.cca.2013.12.042
ISRCTN06414473; Zimmerman et al., submitted) by administration of dehydroepiandrosterone (DHEA) and, as expected, revealed lower total and free testosterone levels in the women taking COCs. It has also been known for a long time that measuring testosterone in sera from women presents a huge challenge to the laboratory, especially if the levels are low [2,3]. The need to introduce better testosterone assays has been emphasized by Rosner et al. [4]. In response to this broadly accepted appeal laboratories have introduced better assays and in this respect liquid chromatography tandem mass spectrometry (LC–MSMS) methods have been considered as the method of choice by a number of laboratories [5,6]. Similarly, manufacturers of immunoassays have made an effort to improve their assays [7]. A new reference management procedure for total testosterone has been proposed [8] and the Center for Disease Control has initiated a testosterone standardization project (http://www.cdc.gov/labstandards/pdf/hs/HoSt_ protocol.pdf, accessed August 2013). In the abovementioned trial on the restoration of testosterone levels during oral contraceptive use total testosterone was measured with a
A.C. Heijboer et al. / Clinica Chimica Acta 430 (2014) 92–95
direct competitive immunoassay and in view of the prevailing concern regarding the quality of testosterone assays in women, the present experiments were designed with the aim to confirm the data obtained with this immunoassay by LC–MSMS. 2. Materials and methods 2.1. Specimens Specimens were derived from a study designed to evaluate the effect of restoring the decreased availably of testosterone with daily administration of DHEA. Design and results of this study, Clinical Trial Registration number ISRCTN06414473, have been submitted for publication. Briefly, the study was a randomized, double blind, placebo-controlled, phase II study performed on 99 healthy new COC users at the University Hospital in Liège, Belgium. The intervention consisted of 3 cycles of a COC containing 30 μg ethinylestradiol (EE) and 3 mg drospirenone (DRSP) followed by 6 cycles in which the COC (Yasmin®) was combined with either 50 mg/d DHEA or placebo. Blood was taken in the morning after an overnight fast at the start of the study and after 3, 6 and 9 cycles of COC use. Serum was stored at −80 °C until analysis. 2.2. Testosterone and SHBG assays As part of the clinical study, testosterone levels were assessed by the Immuchem Double Antibody Testosterone 125I RIA kit according to the instructions of the manufacturer (ICN Biomedicals, MP Biomedicals, The Netherlands, RIAICN). Cross-reactivity with DHEA as reported by the manufacturer was b 0.01%. Two LC–MSMS methods have been applied to selected samples. LC– MSMSQPS was a fast analytical method involving a liquid–liquid extraction step using stable isotope-labeled testosterone (D3-testosterone) as the internal standard. The analytical procedure was as follows: 10 μL of an internal standard solution was added to 250 μL serum sample. After addition of 500 μL ammonium formate (0.1 M, pH 5.0) the mixture was extracted using 3 mL of a hexane–ethylacetate mixture (1:1). The tubes were vortexed and centrifuged for 4 min under ambient conditions at 1500 ×g after which the organic layer was separated and evaporated at 45 °C under a gentle stream of nitrogen. Five microliters was injected into an ultra-high-performance liquid chromatography system (Agilent 1290 series; Hewlett Packard, Palo Alto, CA) equipped with a Sciex API 4000 mass spectrometer (Applied Biosystems, Foster City, CA). The analytical column used was a hydrophilic interaction liquid chromatography column (Kinetex XB-C18 5 × 2.10 mm, inner diameter = 2.6 μm; Phenomenex, Torrance, CA). The following isocratic liquid chromatography solvent program was used: 50% mobile phase A [0.1% formic acid] and 50% mobile phase B [0.1% formic acid in acetonitrile]. The tandem mass spectrometry system was operated in positive ion mode. The retention time of testosterone and the internal standard was b1 min. The following multiple-reaction monitoring transitions were optimized: m/z 289.3 → 96.6 (testosterone) and m/z 292.3 → 96.6 (testosterone-d3). The calibration and quality control samples were prepared in steroid free serum (SeraCare, Milford, MA, USA). The calibration samples (n = 8) ranged from 0.0867 to 86.7 nmol/L (25.0 to 25000 pg/mL) and the quality control samples (at three concentration levels: 0.260 nmol/L (75.0 pg/mL), 4.33 nmol/L (1250 pg/mL) and 69.3 nmol/L (20 000 pg/mL)) were analyzed in duplicate. Intra-assay CV was 6.3%, 4.5%, and 3.3% at 0.26, 4.33 and 69.3 nmol/L (75.0, 1250 and 20000 pg/mL) respectively (n = 8). LC–MSMSVUmc was performed as described before, using D5testosterone as an internal standard [9–11]. Intra-assay CV was 9.5% and 6.0% at 0.21 and 1.98 nmol/L respectively (n = 53). In this method D3-testosterone and D5-testosterone have been shown to be equally efficient as internal standards [12]. Both LC–MSMS methods were shown to be free from interference from DHEA.
93
SHBG in the clinical trial was measured on the Roche Modular E170 platform, SHBGRoche, whereas in the present experiments the Architect immunoassay platform (Abbott diagnostics, Chicago, IL, USA) with an intra-assay CV of 6.3% and 4.9% at concentrations of 34 and 134 nmol/L respectively was used, SHBGAbbott. 3. Results The comparison between the results of testosterone measurements obtained during the trial with the RIAICN and the results obtained with LC–MSMSQPS is displayed in Fig. 1. When analyzed by LC–MSMS, total testosterone is not suppressed after three cycles of COC use. Both testosterone methods showed elevated testosterone levels following addition of DHEA to the COC (Fig. 1). Only in the samples obtained before the use of COC there was concurrence of the results of the testosterone assays. COC use resulted in lower testosterone levels when measured by RIA, irrespective of the addition of DHEA to the treatment. Subsequently, the LC–MSMSQPS was compared to LC–MSMSVUMC in 24 samples with testosterone values ranging from 0.53 to 5.12 nmol/L (by LC–MSMSQPS) selected from three subjects in each treatment arm taken after 0, 3, 6 and 9 treatment cycles respectively. The result (LC– MSMSQPS = 0.931 × LC–MSMSVUmc − 0.059; r = 0.9929; n = 24) establishes the excellent comparability of the two methods and, therefore, confirms the observation that, when measured with LC–MSMS the decrease in circulating testosterone that is observed with RIA is not confirmed. We hypothesized that insufficient release of testosterone from SHBG in the direct RIA procedure might be the cause of this apparent discrepancy. To test this hypothesis, testosterone was measured in 24 surplus serum specimens with a wide range of SHBG levels obtained from young women, 12 of whom were on different COCs. Specimens were analyzed for testosterone by RIAICN and LC–MSMSVUmc. The differences between the results are plotted in Fig. 2 and reveal a bias that is strongly dependent on the SHBGAbbott level (r = 0.79; p b 0.0001). Based on the results of this experiment, we compared the relationship between circulating total testosterone and SHBG using testosterone measurement by RIAICN and LC–MSMSQPS in the specimens from the trial. A statistically significant, negative correlation was found between testosterone measured by RIAICN and SHBGRoche (r = 0.676; p b 0.0001; n = 198). No such relationship was found between testosterone measured by LC–MSMSQPS and SHBGRoche (r = 0.100; NS; n = 196). The results in Fig. 3 confirm our suspicion that the results of the RIAICN are severely influenced by the concentration of SHBG in the specimen. As expected based on the good agreement between the LC–MSMS methods, testosterone measured by LC–MSMSVUmc was not related to SHBGAbbott (r = 0.062; NS; n = 24).
Fig. 1. Serum testosterone before and during treatment with combined oral contraceptives supplemented with DHEA or placebo after cycle# 3, as measured by radioimmunoassay (local RIA) and liquid chromatography–tandem mass spectrometry (LC–MSMS). Data are presented as mean ± SEM (n = 49).
94
A.C. Heijboer et al. / Clinica Chimica Acta 430 (2014) 92–95
Fig. 2. Differences between results of RIAICN and LC–MSMSVUMC for testosterone in relation to the SHBG concentration in the specimen. n = 24; r = 0.79; p b 0.0001.
4. Discussion The present study was designed to confirm the well documented observation [1] that administration of combined oral contraceptives causes a decrease in the circulating level of total testosterone. Worldwide concern about the quality of testosterone assays, especially at low levels, co-fostered our initiative. In sharp contrast to our expectation, however, measurements by LC–MSMS illustrated that after three months of use of 30 μg EE and 3 mg DRSP (Yasmin®), testosterone levels are indiscernible from those prior to the start of the medication. The LC–MSMS method used was compared to another, independent, LC–MSMS method, and these methods were found to agree well. This observation is in sharp contrast to the literature which describes, based on immunoassays with or without extraction [1], a significant decrease in both free and total testosterone during COC use. In contrast to testosterone, SHBG levels sharply rose during COC use in the trial from which the specimens were derived. After three COC cycles SHBG was 241% higher than before treatment and we hypothesized that this increase may have adversely influenced the results of the
Fig. 3. Total testosterone in relation to SHBG in samples from women before and after three cycles of combined oral contraceptive use, as measured by radioimmunoassay (RIAICN; panel A) or liquid chromatography–tandem mass spectrometry (LC–MSMSQPS, panel B).
radioimmunoassay used to assess testosterone levels. Indeed, the difference between the results of the RIA and LC–MSMS was shown to be strongly dependent on the SHBG concentration. This dependency may also explain the differences observed in testosterone levels following addition of DHEA to the COC (Fig. 1). We, therefore, suggest that, when measured with a state-of-the-art method, total testosterone does not change upon oral contraceptive use. Of course, there are differences in the design and quality of immunoassays [4] and not all immunoassays may be equally affected by the SHBG level in the specimens. One way to overcome the potential influence of SHBG on the test result is to extract testosterone with organic solvents prior to the immunoassay. In theory immunoassays preceded by extraction and chromatographic separation of the analyte should yield a result that shows more agreement with LC–MSMS than immunoassays without sample pretreatment. The effect of COC use on total testosterone assessed with extraction based methods should therefore be less than those assessed with direct immunoassays. In fact this was observed in the systematic review [1]. The supplemental material to that paper shows that COC use results in a 0.56 (95% CL 0.50–0.62) nmol/L decrease in serum testosterone when analyzed by direct methods (45 studies on 2098 patients), and 0.37 nmol/L (95% CL 0.26–0.47) when analyzed by methods using extraction and chromatography (17 studies on 522 patients). The difference was found to be statistically significant, p = 0.003 but the observation that even with extraction methods a decrease in total testosterone is observed on meta-analysis requires further consideration. The individual studies collected in the meta-analysis differ with respect to the conclusions. Thirty-five of the 45 studies using direct assays, 78%, report a significant testosterone decrease in the meta-analysis, versus only 9 of the 17 studies with extraction assays (55%). That not all extraction based assays fail to show a significant decrease may have to be attributed to methodological influences and may require further investigation. This may, however, be hindered by the fact that many of these methods may no longer be in use. Interference from the components of the COCs used is considered unlikely, as this would have led to apparently increased testosterone results in the competitive immunoassays and not to a decrease. Similarly, insufficient extraction recovery is also not likely to explain the phenomenon as this would equally affect specimens from control and COC cycles. The only exception to this would be if the increased SHBG in the specimens obtained during COC use would cause a difference in extraction efficiency. Dependency on the concentration of binding proteins is not unique to testosterone, since it has also been reported for vitamin D [13]. Unfortunately, like immunoassays, LC–MSMS methods also do not all yield the same result, especially at low levels [14,15] and with some LC–MSMS methods, total testosterone has been reported to decrease modestly during oral contraceptive use [16]. Moreover, these authors report on age-specific reference values tor testosterone in women, whereas we have not taken possible age differences into account. As in our study however, free testosterone was much more suppressed than total testosterone. Based on their results on postmenopausal women, Legro and colleagues [15] claim that the radioimmunoassays, when properly maintained, should yield comparable results to LC–MSMS methods. This may be true, but in postmenopause SHBG levels are not that high compared to after use of some oral contraceptives. Provided that their RIA can be shown to be free from the SHBG interference as described here, we agree with their conclusion. Our finding that, when measured by LC–MSMS, total testosterone, in contrast to earlier studies, does not change during COC use has no impact on the notion that women on COCs may experience androgen deficiency. The substantial increase in SHBG concentration during the use of many COCs will lead to a sharply decreased free testosterone level which may be responsible for the adverse effects reported. Studies on the benefit of maintaining bioavailable testosterone to the pre-COC level will, therefore, have to include assessment of SHBG and a proper method of measuring or calculating free testosterone. Preferred
A.C. Heijboer et al. / Clinica Chimica Acta 430 (2014) 92–95
methods are the mass action equation and equilibrium dialysis [3], the latter being quite laborious though. In summary then, in contrast to a vast number of reports in the literature, our results support the hypothesis that total testosterone concentrations do not change during the use of combined oral contraceptives when measured with a state-of-the-art method. The discrepancy with the literature can most likely be explained by the sensitivity of immunoassays to SHBG, as we showed strong dependency on SHBG concentrations in a testosterone immunoassay. Obviously, our data are limited to one immunoassay and the SHBG dependency of other immunoassays should be documented. Acknowledgments The authors gratefully acknowledge the expert technical assistance by F. Martens (VUmc) and B. Ottjes (QPS). References [1] Zimmerman Y, Eijkemans MJC, Coelingh Bennink HJT, Blankenstein MA, Fauser BCJM. The effect of combined oral contraception on testosterone levels in healthy women: a systematic review and meta-analysis. Hum Reprod Updat 2014;20:76–105. [2] Taieb J, Mathian B, Millot F, et al. Testosterone measured by 10 immunoassays and by isotope-dilution gas chromatography–mass spectrometry in sera from 116 men, women, and children. Clin Chem 2003;49:1381–95. [3] Miller KK, Rosner W, Lee H, et al. Measurement of free testosterone in normal women and women with androgen deficiency: comparison of methods. J Clin Endocrinol Metab 2004;89:525–33. [4] Rosner W, Auchus RJ, Azziz R, Sluss PM, Raff H. Position statement: utility, limitations, and pitfalls in measuring testosterone: an Endocrine Society position statement. J Clin Endocrinol Metab 2007;92:405–13.
95
[5] Soldin SJ, Soldin OP. Steroid hormone analysis by tandem mass spectrometry. Clin Chem 2007;55:1061–6. [6] Stanczyk FZ, Clarke NJ. Advantages and challenges of mass spectrometry assays for steroid hormones. J Steroid Biochem Mol Biol 2010;121:491–5. [7] Groenestege WM, Bui HN, ten Kate J, et al. Accuracy of first and second generation testosterone assays and improvement through sample extraction. Clin Chem 2012;58:1154–6. [8] Botelho JC, Shacklady C, Cooper HC, et al. Isotope-dilution liquid chromatography– tandem mass spectrometry candidate reference method for total testosterone in human serum. Clin Chem 2013;59:372–80. [9] Bui HN, Struys EA, Martens F, et al. Serum testosterone levels measured by isotope dilution-liquid chromatography–tandem mass spectrometry in postmenopausal women versus those in women who underwent bilateral oophorectomy. Ann Clin Biochem 2010;47:248–52. [10] Bui HN, Sluss PM, Blincko S, Knol DL, Blankenstein MA, Heijboer AC. Dynamics of serum testosterone during the menstrual cycle evaluated by daily measurements with an ID-LC–MS/MS method and a 2nd generation automated immunoassay. Steroids 2013;78:96–101. [11] Bui HN, Schagen SE, Klink DT, de Waal HA Delemarre-van, Blankenstein MA, Heijboer AC. Salivary testosterone in female-to-male transgender adolescents during treatment with intra-muscular injectable testosterone esters. Steroids 2013;78:91–5. [12] Bui HN, Blankenstein MA, Heijboer AC. Isotopically labelled testosterone derivatives as internal standards in liquid chromatography–tandem mass spectrometry. Ann Clin Biochem 2013;50:275–6. [13] Heijboer AC, Blankenstein MA, Kema IP, Buijs MM. Accuracy of 6 routine 25hydroxyvitamin D assays: influence of vitamin D binding protein concentration. Clin Chem 2012;58:543–8. [14] Thienpont LM, Van Uytfanghe K, Blincko S, et al. State-of-the-art of serum testosterone measurement by isotope dilution-liquid chromatography–tandem mass spectrometry. Clin Chem 2008;54:1290–7. [15] Legro RS, Schlaff WD, Diamond MP, et al. Total testosterone assays in women with polycystic ovary syndrome: precision and correlation with hirsutism. J Clin Endocrinol Metab 2010;95:5305–13. [16] Haring R, Hannemann A, John U, et al. Age-specific reference ranges for serum testosterone and androstenedione concentrations in women measured by liquid chromatography–tandem mass spectrometry. J Clin Endocrinol Metab 2012;97:408–15.
